--- In cybalist@..., "ravichaudhary2000" <ravi9@...> wrote:
> --- In cybalist@..., "richardwordingham" <richard.wordingham@...>
> wrote:
> d see if there is any rational basis to these
> > > papers.
> > > Do the mathematical concepts make any sense?.
> >
> > No.
> >
> > No. There are 230 3-D symmetry groups
> > (http://www.iucr.org/iucr-
> top/comm/cteach/pamphlets/21/node4.html).
> >
> > 230 <> 17 x 17.
> > Can we set up sensible features to get the following grid?
> >
> > k kh g gh ng h. h a
> > c ch j jh n~ s' y i
> > t. t.h d. d.h n. s. r r.
> > t th d dh n s l l.
> > p ph b bh m w u
> >
> > I can't see any useful way of fitting in m. e o ai au.
> >
> > Richard.
>
>
>
> My reasons for posting the abstract, on this technical linguistic
> list, was that here is someone attempting to explain a language
> construct with a mathematical model. which in itself is admittedly
> not a very common approach. I am hoping that some with more
knowledge
> than I have, could analyze , criticize, and develop the subject
> further.
>
> Mr Wordingham, rightly, applies his critical faculties, to the
model,
> and raises some issues, with examples from other languages, and the
> English language for example has 23, 24, 25, or 26 characters(
from
> a mathematical perspective, we now have a model, with a range of
> variables from 23 to 26 this is a mathematical model, albeit
a
> rudimentary one).
Actually, 25 was the one letter count I couldn't derive!
> That forms the basis of a scientific discussion, and I hope others
> will be expand on his comments, and develop the hypothesis further.
It is generally argued that, in an ideal world, the number of letters
for a language would correspond to the number of phonemes. There are
a few pragmatic issues that commonly complicate even this simple
scheme. Some contrastive features may be represented by an extra
symbol, e.g. vowel length, nasalisation, aspiration and consonant
doubling (which I suppose one might want to consider lengthening),
and this disrupts the patterns. Moreover, the consonants and vowels
rarely pattern together as neatly as in Vedic Sanskrit,
with its 5 series of 9 orders (voiceless, voiceless aspirate, voiced,
voiced apsirate, nasal, fricative, semivowel, short vowel, long
vowel) and then 4 diphthongs left over. (I dimly recall there being
an /f/-like glide in Vedic Sanskrit.) These phonemes are better
tallied up as 6x5 + 4 (yrwl) + 1 (f) + 2x5 + 2x2, and the inclusion
of f and h. with the sibilants to make a 'fricative' order is still
frankly dubious. It is a case of forcing a pattern where there isn't
really one. The traditional arrangement is probably better still.
(One can tidy up the pattern further by eliminating the diphthongs by
treating [e] and [ai] as /ay/ and /a:y/!) While the number of
phonemes in a language may be meaningfully broken up in this sort of
pattern, the overall total is rarely elegant. Moreover, gaps in
these patterns are not unknown.
I ignored anusvara (m.) from the above; one can regard it as a
variant of /m/ or treat it as nasality on the vowels.
Incidentally, I cheated when counting the phonemes of Vedic
Sanskrit. My understanding is that the Brahmi alphabet was developed
for an early Prakrit, and then adopted by Sanskrit. This explains
why there is a letter for /n~/ although it does not contrast with
guttural /N/. I wonder how genuine the retroflexes are for Sanskrit
as a mother tongue; could they be borrowings from early Prakrit? I
think there may be an interesting story in the origin of Indian
retroflexes. Does Dr Kalyanaraman know a good on-line account?
(/s./ has a sound IE origin, but how did /n./ come to contrast
with /n/? How did the retroflex plosives of Sanskrit originate?
Surely not from /s.t/ > [s.t.]! I know that in generally retroflex-
free Europe Swedish and one Yorkshire dialect have developed
retroflexes from /r/ + dental plosive.)
The number of letters becomes even more arbitrary when an alphabet
has a long history. Even the 26-letter Roman alphabet has primarily
redundant letters (C or K, Q, X and ultimately Y), but some of these
have subsequently been put to good use in some languages.
The 32 consonants of Persian include many which are redundant other
than to give Arabic loan words the same spelling in both languages.
> Sanskrit, if I understood right, is a structured, indeed, some
think
> an artificial language. Could the rules, not be analogical to
> mathematical rules? The grammarian Panini's, works have been
studied,
> and some have proposed that here was a pre runner to today's
database
> systems, indexing, and the strings that are used in computer
> programming.
There is nothing new under the sun.
> Structures require rules. Structures lend themselves to
mathematical
> models. If languages have structures, there is then nothing, from a
> conceptual basis, that says that mathematical modeling techniques
> cannot be applied.
True, *if* the rules have been captured.
I do remember Piotr writing that mathematical attempts to deduce and
apply sound changes, e.g. for lexicostatistics, had been very
disappointing. Also, I suspect that pleasant numerical patterns do
not appear.
A popular scheme has been to reduce _everything_ to binary
oppositions. However, the results frequently appear unnatural.
However, all is not lost. I would presume that Babelfish uses rules.
> Mr Wordingham is also right in not applying one standard to other
> papers on the waves list, over 120, covering a vast range of
> subjects, from Philosophy, archeology medicine, the epics, culture
> and so on.
>
> As for the paper itself, what is available is just an abstract.
Maybe
> the author has some rational basis, for his concepts. The author
may
> also be totally wrong.
>
> With the information available, we just do not know.
>
> Again, I suggest, Richard could contact, the author, his address
is
> available, obtain the paper, and give us the benefit of his
> considered analysis.
The primary purpose of the abstract of a paper is to enable potential
readers to decide whether they should read the paper. It may
secondarily serve as a summary for those who cannot read the paper.
Unless someone, e.g. Dr Kalyanaraman, who was at the conference,
knows better, I will assume that the abstract gave a good indication
of the contents. I see no reason to put myself under an obligation
just to receive a pile of mystic numerology.
Richard.